Maximum-Likelihood Prediction and Estimation for Nonlinear Dynamic Systems

摘要

A method is given for determining the system state using noise-corrupted observations of a non-linear dynamic invector process, with a numerical application to radar observation of a reentry body. The study examined the feasibility of numerically solving the vector-differential equations satisfied by the maximum-likelihood estimator. The maximum-likelihood estimate is that initial condition which minimizes a certain functional on itself, on the observation, and on the a priori statistics. (Author)